Genetic variation is a hallmark of retrovirus infection and occurs in avian, murine and human retroviruses such as HIV. The major focus of this project is the study of genetic variation in retroviruses and the effect of such variation on virus infection and disease. Studies currently underway in the laboratory include the determination of the point mutation rate of retroviruses; the polymorphism of recombinant host range retrovirus variants; pseudotyping as a result of mixed retrovirus infections; and the treatment of retrovirus infections with immunotoxins reactive with retrovirus-coded antigens. A. Reported estimates of mutation rates of retroviruses suggest that progeny retroviruses sustain several point mutations during a replication cycle. In contrast we have determined that most progeny murine retroviruses do not sustain any point mutations during a single cycle. Studies are underway to compare the point mutation rate of viable progeny viruses to that of proviruses in the absence of selection for viability. B. Inoculation of many murine retroviruses in mice results in the generation of host range variants by recombination of the inoculated virus with endogenous retroviral genes. We have identified two major antigenic subgroups of these variants. The subgroups appear to be derived from distinct endogenous sequences and differ with respect to their in vitro cell tropisms. We are currently constructing chimeric viruses between viruses of the two major subgroups to map the genetic determinants of their properties. C. The generation of host range variants in retrovirus-infected mice results in a mixed retrovirus infection which may give rise to viral pseudotyping. We have found this to be the case in mice infected with the Moloney lymphocytic leukemia virus. Moreover, the pseudotyping appears to be a dynamic phenomenon which varies in different organs, changes during the course of disease and may influence the spread of infection of the target organ by host range variants. Continuing studies will examine the extent of virus spread and the cell-types infected by the host range variants. D. Ricin A-chain immunotoxins constructed using antibodies reactive with murine retrovirus-coded proteins have been tested for efficacy in selectively killing infected cells. Several immunotoxins have been found to be effective using infected in vitro cell lines. Initial studies using a murine retrovirus-induced neurodegenerative disease as a therapeutic model suggest that certain of the immunotoxins are also cytotoxic to infected cells in vivo. A retrovirus-induced lymphocytic leukemia is currently being evaluated as a second in vivo therapeutic model.